Modeling and Analysis of Social Communication Networks
(selected publications)
We are
interested in
understanding how laws governing individual behavior (micro-laws)
influences the
macroscopic communication structure of a large society. The inference
of the micro-laws from macroscopic history then gives
insight into the behavior of the society, both in the past and
future, as well as the
ability to identify comunication structures that may have been hidden.
Modeling, Simulation, Prediction and Control
We have developed a general model that incorporates a wide range of
possible behaviors. This model can be simulated given a current society
for the purposes of prediction. We are currently interested in the
stochastic control problem of determining what the laws of the
society should be givne a certain desired future
state.
Identification of Hidden Structure
We are interested in using communication data of societies (for example
newsgroup postings) to identify whether an as yet undiscovered
group is operating within the society.
Reverse Engineering Society Laws
Given the observed history of a society, we use machine learning methods
to determine the laws of the society that could have given rise to that
behavior. These reverse engineered laws can then be used for prediction.
Informtaion Retrieval From (Social) Networks.
How to retrieve and visualize information pertaining to networks,
such as clusters, structure, importance of nodes and groups, etc.